So what is the learning here, why did it fail? From the timeline it looks like it spread out too much and didn‘t create a great product first.
Comments from insiders in Swedish[1] and Dutch[2]. Some snippets: ... In the end, they are just in a situation that is almost impossible to save. You have a factory full of machines that are substandard in quality, reliability and documentation. A huge 100% in-house tech stack that largely consists of Go pieces on Lambdas writing to DynamoDB. ... ... A gigantic factory full of mediocre Chinese equipment, what can you…
> In theory it's microservices, but the reality is that there are so many circular dependencies that it works like a monolith
But lambda/go/dynamodb does not force this situation.